Experimental validation of a high‐gain observer for composition estimation in an ethanol–water distillation column

Abstract

In this paper a high-gain observer used to estimate the product compositions in a distillation column for a non-ideal mixture (ethanol-water) through the tray temperature measurements is presented. The design of this observer is based on a simplified mathematical model. One of the main advantages of this observer is its constant gain, therefore its tuning depends only on choosing a few constant parameters satisfying some simple algebraic inequalities. The effectiveness of the proposed method is demonstrated through on-line experiments in a distillation pilot plant. 

Control Tolerante a Fallas Activo en Sensores y Actuadores: Aplicación a una Columna de Destilación

“…El observador de alta ganancia empleado fue propuesto en (Hammouri et al, 2002;Targui et al, 2002) con resultados favorables a nivel experimentación (Téllez-Anguiano et al, 2010). Para el diseño del observador de alta ganancia se considera un modelo no lineal de la forma dada en (12), con un sólo estado medido ( ) y un vector de perturbación (d):…”

“…Dada la importancia de la seguridad para los usuarios y equipos de procesos industriales nuestro objetivo es desarrollar un sistema de CTFA para una columna de destilación, donde la principal aportación y que lo hace diferente de otros trabajos propuestos (Flores, 2012;Aguilera, 2012;Téllez-Anguiano et al, 2010) es el esquema para detectar y diagnosticar fallas múltiples y s en sensores y actuadores, sin considerar un banco de observadores para detectar fallas en sensores. Además, el modelo de la columna de destilación que se empleó para el desarrrollo de los observadores de alta ganancia y adaptativos considera el método denominado NRTL (por su acrónimo en inglés, Non Random Two Liquid) para el cálculo de la concentración.…”

Singular linear parameter‐varying observer for composition estimation in a binary distillation column

“…The estimation problem in distillation columns has been discussed in several works, for example, [27][28][29][30][31]. In this work: (i) the problem consists of designing a PI-Observer using the singular LPV model (25), in order to estimate the compositions of the mixture (ethanol-water), and (ii) we are interested in validating the proposed observer through experimental data.…”

“…where j(t) [ R n is the state vector; u(t) [ R m with (m < n) is the control input vector, and @(t) [ R p is the measurement output vector; d(t) [ R l is the unknown bounded process disturbance vector; E [ R n×n is a singular matrix with rank (E) < n and g(•) is a continuous non-linear function and infinitely differentiable. Taking into account the dynamics of the plant, system (29) can be rewritten in the form (30)…”

Composite desirability function‐based empirical modeling for packed tower design in physical ammonia absorption

“…Although there are several other successful studies in the recent literature on the packed columns, distillation columns and packed media, there are almost no systematic papers regarding the implementation of different mathematical analogies leading to the formulation of novel prediction models, which can directly be used for the packed tower design in physical ammonia absorption. To the best of the author's knowledge, the present paper is the first study specifically devoted to a study of composite desirability function‐based empirical modeling for the prediction of packed column diameter and packing height in ammonia absorption.…”